Key Company Stats:
9 years
Founded in August 2013
200+
Employees across 8 hubs
28 languages
Marketing + Customer Service translations
1M+
CS messages translated monthly
Background:
Delivering a first-class customer experience matters more than ever — it’s been proven to propel revenue growth and contribute to wider business objectives. For fast-moving global brands, this means continually engaging new and existing customers with great-quality content in their local language. But as any scaling organization knows, the challenges can ramp up quickly, as marketers continue to add more content and languages as they enter new markets.
In July 2022, AI-powered, human-refined translation platform, Unbabel, sought to survey a key persona — 1,668 global marketers across the US, UK, Japan, Brazil, France, Germany, Spain, and Sweden — to understand the challenges and successes it was having with localizing content at scale, as well as to uncover any key trends.
The research would be compiled for its annual market research report: Global Marketing Localization Trends 2023, providing other marketers access to insights about wider trends in international marketing and localization directly from their global peers, including those from non-English speaking countries.
The Challenge:
The marketing team tasked with writing the questionnaire and fielding the survey are English native speakers, who needed to prepare the survey in six additional languages by particular locales — Japanese, Brazilian Portuguese, German, European Spanish, French (France), Swedish — plus one other that required British English. The survey included logic, as well as questions ranging from single select, rank, sliding scale, and open-ended, which all needed to be QA’d, launched, and fielded within a 3-week turnaround time.
As a global organization, with headquarters in Lisbon, Portugal, Unbabel houses a slew of multilingual talent from all over the world, including some native to key languages that were being targeted, such as Portuguese. And while 36% of marketers rely on internal colleagues for whom translation is often not their job, employing internal talent who are not professionally trained translators can compromise speed, quality, and accuracy. Because even though they share the same language, Portugal and Brazil have vastly different meanings for some words and phrases, which can lead to irreparable misunderstandings.
And word-for-word textbook translations that generic MT spits out were simply not a viable business solution, due to the nuanced nature of some of the survey questions, plus lack of human refinement.
With a survey incidence rate of 50% and high impact to overall feasibility, Unbabel needed a reliable language operations solution to help it achieve translation at speed and quality. And it didn’t have to look far.
The Unbabel Approach:
Unbabel’s project management team assigned translators from their very own global community of 100,000+ enrolled multilingual editors — composed of native speakers, linguists, subject matter experts, and professional translators — working across 130+ language pairs.
When all the translation files were successfully uploaded back into SurveyMonkey, Unbabel assisted with checking the survey links to ensure the logic worked properly and that the survey went through to completion as expected. It also ensured the correctness of the translations and fixed the entities that were not being parsed correctly, resulting in unnecessary code.
Where standard machine translation lacks human-quality elements, Unbabel relied on the expertise of its own AI-powered tech combined with its proprietary human translation network to help spot errors that can simply be overlooked: In this case, the Japanese editor was able to flag an ampersand symbol that due to SurveyMonkey limitations, did not properly render during his QA of the survey.
Once all the replies were collected, Unbabel facilitated translating both the open-ended and the “other, please specify” replies back into English, enlisting the help of its Localization Engineering department, as these files needed pre- and post-filtering.
Results:
Unbabel was able to successfully achieve 111% survey completion rate within a 2-week turnaround time — 1 week ahead of schedule. It collected compelling responses from team lead to C-suite level, across a wide range of industries including retail & consumer goods, software & tech, and healthcare & pharmaceutical, exceeding response expectations in some key markets (~+20%).
Have an upcoming market research project? See how Unbabel can help field yours!
